Zabbix Conference 2014 Agenda

Presentation on how ChinaNetCloud uses Zabbix to monitor thousands of hosts in over 100 datacenters and millions of items. Includes architecture, lessons, GUI customizations, custom scripts, templates, security, and how we deeply monitor the world's Internet servers serving hundreds of millions of users. I'll also discuss how we track the dynamic nature of Zabbix, where we perform auto-provision, auto-discover, and cross-check a system, where we add, remove and re-configure many hosts every day.

Big companies have big issues and uncommon constraints because of the number of components and the complexity of human organization managing the servers and the security, specially in the banking sector. As a consequence, it implies strong constraints on Zabbix platform in terms of sizing, architecture, monitoring organization, users grants, performance, etc. In such environment, it is common to have 10 different teams working on the same server, each team wants to deal only with their own alarms, filtering the events by team is really a deal. The Zabbix monitoring is managed by over 100 Zabbix admins, avoiding the overlappings is another one. Deployment in those companies is done from a Component Management Database, 500 servers configuration changes per day is common, Zabbix has to be updated immediately/automatically. These are just some of the issues of many others, we will see how they have been solved in 2 of the top 5 French Banks.

Zabbix can be used not only for listening and alerting, but also as an active control system. For that purpose the monitoring process must be proactive, because the control reaction to the alert is usually not fast enough and needs a certain time for execution. This paper presents the approach and real-world examples on how Zabbix is used to predict and prevent the anomalies beforehand. Co-author of the presentation: Dmitry Shchemelinin, Ph.D., Sr. Director of Operations, RingCentral Inc.

The topic will cover general advice for MySQL tuning as well as 5.6 specific tunables

There are many ways to bring the power of JIRA and Zabbix together. This talk will focus on some of them, covering creating JIRA tickets from Zabbix events and monitoring JIRA ticket queues by Zabbix.

This talk is about Monitoring Storage Area Network for a big Zabbix customer in Italy. Scenario over 150 Brocade switch with over 200 ports per switch - 1200 NVS. During our Zabbix monitoring implementation we have found a lot of trouble fixed in collaboration with Zabbix Team.

From being the "wild, wild, west" or a purely academic and entertaining medium, internet has become a daily, vital part of our life. And the internet radically changed in late 2013, when ICANN started introducing hundreds of new Top Level Domains into the root zone, as part of the new generic Top-Level Domain Program. Hundreds of new TLDs mean hundreds of new opportunities for communities and organizations to express themselves, in ways never imagined before. I will shortly explain how the DNS industry work, from a technical perspective, as well as how ICANN is using the Zabbix monitoring system and Zabbix SIA consulting services to monitor the internet's domain name system.

I have written 2 modules by Ansible to install the Zabbix-Server and the Zabbix-client. With the Zabbix-server module it is possible to install the server in only a few minutes with predefined settings for the Zabbix server, database and webserver even when using a distributed setup. The Zabbix-client module will do the same for the client. It will install the Ansible client on 1 or multiple hosts at the same time and modify the Zabbix configuration so that no manual adjustment of the zabbix_agentd.conf file is necessary.

ZabbixUnify is a multi zabbix frontend, written in python using Pyramid and Twitter Bootstrap frontend frameworks, the idea of using Twitter Bootstrap came from the need of having a user friendly mobile interface. ZabbixUnify works by connecting to multiple Zabbix frontends using APIs, all zabbix versions (1.8+) are supported, it asynchronously gets information about hosts and triggers from all the configured zabbix frontends and shows them in a single interface, it's then possible to drill-down to see information about all the active triggers for all hosts of a single Zabbix installation, and it's also possible to show detailed information about active triggers like: name, latest values and date.

More or less Zabbix is a database-agnostic system. So you may run it, for example, on the Oracle RDBMS if you have a spare license. And here it is how to get the most out of it, what problems you may meet and how to deal with them.

This talk will cover information about: - current monitoring configuration process and its disadvantages and error-proneness - requirements for a process automation and scalability - Zabbix API as a proper solution - implementation within a custom management tool - how we tackled performance issues in a steady growing platform - experiences and outcome of the project - optional: experiences with migration from Zabbix 1.8 to 2.2

In this talk Alexander will discuss details of available Zabbix Professional Services and the benefits your company/organization gets when using these.

We have a lot of information inside of each monitoring server. It is possible to delete old history and not to take care about it or to analyze it. This project tries the second option. There can be very nice abnormalies or incidents, which are hidden if history is deleted and not analyzed. Nowadays network technologies can handle very high transfer speeds. It is still more difficult to do network flow analysis because of high speed and the needs of more powerful hardware. Current methods are based mainly on the analyze of data flows and patterns. It is quite impossible to find abnormalities on high speed link, because small change in flows can mean big excesses in incidents. Thesis project focuses on multi criteria analysis of data to be able to predict incidents. Due to this reason system monitoring seems to be good core for it.

Come to this talk if you want to better understand how a multi-core multi-disk system operates and how you can inspect its operation on Linux in order to measure utilization for capacity planning, or just for fun. It has been long time ago, when a commodity server had one single-core CPU, one disk and understanding of a system utilization was as easy as checking that the load average is below 1. Yet some may still think of a system as highly utilized if not saturated when the system load average goes above and beyond 2-3. What does the load average really mean on a multi-core multi-disk server? Why are my tasks still slow if CPU is never more than 13% utilized? How much more disk resources do I have? Why can't my application read data any faster even though we have 50 disk array now? After this talk you will find answers to these and many similar questions.

Understanding performance tuning is an important subject for all systems administrators. Come learn more about what it actually means to measure performance. In addition Andrew will share his experiences of how to ensure your next performance tuning project is successful. My intention is to take a standard application/service such as Apache (used to run websites) and show how to visualize peak performance, especially since out of the box this is not very easy to do. In addition I would go into discussing the process of what metrics are useful to measure in helping to improve the performance of the application. And yes, almost all of this will be done using Zabbix as the measurement tool.

Monitoring process is a particular part of IT. Most of metrics collected by monitoring solutions is just a representation of low-level OS/application statistics. When we speak about monitoring, for example, application response - the common way is creating some kind of synthetic responses and say: "Well, my app have no problems 'cos test response is in range of n milliseconds". But there are real user-side response time (and many more metrics) in app logs. This presentation presents you a powerful approach to make a log-driven monitoring infrastructure. Testbed: approx. 50 servers with erlang-based Zabbix client & log-driven approach.

All about the internals of the Zabbix database, tables, statuses, relationships, and how we use SQL every day to cross-check data, find missing elements, answer questions, track the queue, do housekeeping, special auto discovery, and much more. I will cover the 10-20 reports we use daily to validate our large Zabbix system, which has new hosts and configuration changes every day. I will also describe how we created a new useful audit system to find who changed what and when with lookups by common elements. Plus, how we extended the DB a little to integrate into our ticket and other real-time systems. And, Zabbix 2.2 ER diagram from MySQL WorkBench.

An overview of the methods available for monitoring of existing logfiles, and creating virtual logfiles within Zabbix.

Zabbix is a great monitoring platform, however, there are some visible drawbacks as well. We may bring Zabbix to a new level of functionality, usability and scalability provided we fix a number of the most annoying issues. That's what my talk is about.

During this session all participants and attendees of Zabbix Conference 2014 are welcome to ask their quesions to Zabbix team. Zabbix team members will be ready to answer all your questions regarding Zabbix products, services and future plans. Rihards_Olups Moderator: Rihards Olups